Google and Microsoft are restricting their search APIs, with Bing retiring its API and Google limiting results to 10 per query. This shift reflects their strategy to control web data access through AI-mediated services within their cloud ecosystems rather than open retrieval. The restrictions create opportunities for new players like Perplexity and Parallel, who are building AI-native search APIs optimized for retrieval-augmented generation (RAG) workloads. These emerging platforms emphasize retrieval quality, low latency, and open access, often built on infrastructure like Vespa. The change marks a transition from raw result APIs to search infrastructure integrated directly into AI pipelines.
Sort: